BingoMini I Technical White Paper

1. Core Processing Unit

World-leading embedded AI compute platform for stable, sustained local large model inference.

• SoC: NVIDIA Jetson AGX Orin
• AI Performance: 275 TOPS (INT8)
• Memory / VRAM: 64GB LPDDR5 (large LLM and complex vision algorithms)
• GPU Architecture: NVIDIA Ampere GPU, 2048 CUDA cores, 64 Tensor cores

2. Storage & Data Management

High-capacity, high-throughput local storage for personal and family privacy.

• System / App Storage: 1TB NVMe M.2 SSD
• Data Sovereignty: All interaction data, generated content, and documents stay in local 1TB space. Physically isolated from cloud breach risk.
• I/O Performance: High-bandwidth NVMe for smooth local knowledge base access during AI tasks

3. Physical Specification & Interaction Design

Industrial aesthetics with extreme ease of use.

• Dimensions: 15 cm × 15 cm × 20 cm
• Enclosure: Cyber-style black metal / high-performance polymer, efficient cooling grill
• Display: 4.x" embedded touchscreen. Real-time IP display, status, one-touch setup
• Power: ≤ 60W (low standby, 24/7 ready)

4. Connectivity & Interface

• Wireless: Dual-band Wi-Fi
• Wired: 1000M LAN (RJ45)
• Setup: No keyboard, mouse, or monitor. Scan QR code in mobile browser to connect

5. Environmental & Compliance

• Use Case: Home, small team (recommended ≤3 concurrent users)
• Operation: Fully offline AI on pure local LAN